[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <48B82843.4090707@gmail.com>
Date: Fri, 29 Aug 2008 09:48:03 -0700
From: Tom Quetchenbach <virtualphtn@...il.com>
To: Denys Fedoryshchenko <denys@...p.net.lb>
CC: David Miller <davem@...emloft.net>, netdev@...r.kernel.org,
lachlan.andrew@...il.com
Subject: Re: [PATCH][BUG] tcp: advertise MSS requested by user
Best I can tell, my mailer ate this reply, so apologies if you get it twice.
>> Thanks for your patch, I'll mull over this and put in the fix
>> once I understand exactly how this got broken.
David, thanks a lot for looking into this. It will be really helpful for
our TCP benchmarking work.
> I am using, just i workaround this over iptables MSS option. You feel
> discomfort, but most people start googling and see this trick with iptables,
> and they use it.
In my case I don't think the iptables workaround is the right solution.
I'm working on a traffic generator (tmix, from the University of North
Carolina), that starts many simultaneous flows between two endpoints,
based on a set of "connection vectors" derived from a tcpdump. The MSS
for each flow is specified in the connection vector file on a per-flow
basis. So it's not so simple as clamping the MSS to the path MTU or
setting it based on some simple criteria like destination or port
number. I'm sure I could hack something together using iptables to
approximate what I want to do, but it seems like a bit of a mess to me.
I'm not that familiar with iptables so please correct me if I'm wrong.
Thanks
-Tom
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists